INDEX

Explanations

quantitative comparisons or changes

oai_token-act-pair · gpt-3.5-turbo

phrases indicating changes in quantity or degree

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

umbn

-0.71

 Corpus

-0.69

opus

-0.65

agher

-0.62

 dedication

-0.61

erity

-0.58

çīĪ

-0.58

olitan

-0.56

ivia

-0.55

arro

-0.54

POSITIVE LOGITS

 notch

0.91

 precip

0.89

 fold

0.82

fold

0.82

 steep

0.78

 percept

0.75

 sharply

0.73

astically

0.72

dB

0.72

 fortunes

0.72

Activations Density 0.175%

quantitative comparisons or changes

phrases indicating changes in quantity or degree

No Comments

No Known Activations