INDEX

Explanations

fragments related to reading or watching something, with a preference for verbs associated with consumption like "reading" and "watching."

oai_token-act-pair · gpt-3.5-turbo

mentions of media content, specifically watching or reading experiences

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

borg

-0.69

allas

-0.69

CFR

-0.67

 Americas

-0.66

©¶æ¥µ

-0.66

FSA

-0.65

¬¼

-0.64

 Caribbean

-0.63

pi

-0.61

 metals

-0.60

POSITIVE LOGITS

 aloud

1.72

 myself

0.95

 religiously

0.88

uscript

0.87

 honestly

0.86

 anyways

0.86

 objectively

0.85

 orally

0.85

 beforehand

0.83

 faithfully

0.81

Activations Density 0.283%

fragments related to reading or watching something, with a preference for verbs associated with consumption like "reading" and "watching."

mentions of media content, specifically watching or reading experiences

No Comments

No Known Activations