INDEX

Explanations

punctuation marks which appear in various contexts

oai_token-act-pair · gpt-3.5-turbo

repetitive phrases or expressions often characterized by the word "no"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

gypt

-0.74

 chains

-0.66

 herds

-0.61

ounded

-0.60

vez

-0.60

ãĤ¼ãĤ¦ãĤ¹

-0.59

İĭ

-0.59

flows

-0.58

 feasibility

-0.57

mud

-0.57

POSITIVE LOGITS

 whatsoever

1.00

nor

0.93

sir

0.86

 except

0.79

 please

0.71

uh

0.70

 thank

0.69

please

0.67

elsius

0.65

opes

0.62

Activations Density 0.071%

punctuation marks which appear in various contexts

repetitive phrases or expressions often characterized by the word "no"

No Comments

No Known Activations