INDEX

Explanations

numbers or numerical expressions

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, specifically periods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 proactive

-0.72

 burst

-0.66

 unders

-0.64

 posture

-0.63

 blush

-0.63

 pooled

-0.63

 butcher

-0.63

bay

-0.62

 informed

-0.62

 active

-0.61

POSITIVE LOGITS

1.16

0.98

 Additionally

0.97

0.96

sbm

0.95

0.92

0.91

 Also

0.90

0000000000000000

0.89

0.88

Activations Density 0.200%

numbers or numerical expressions

punctuation marks, specifically periods

No Comments

No Known Activations