INDEX

Explanations

patterns indicating important information or signals

oai_token-act-pair · gpt-3.5-turbo

phrases that suggest inference or indication of conclusions

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

@#&

-0.79

fare

-0.77

jan

-0.73

zan

-0.69

assic

-0.69

zanne

-0.68

aders

-0.66

ÄŁ

-0.65

ITS

-0.64

quer

-0.64

POSITIVE LOGITS

ively

0.95

ered

0.85

ially

0.79

 otherwise

0.76

 indications

0.74

 displeasure

0.71

iveness

0.71

icity

0.70

 willingness

0.70

 signs

0.68

Activations Density 0.052%

patterns indicating important information or signals

phrases that suggest inference or indication of conclusions

No Comments

No Known Activations