INDEX

Explanations

instances where something was previously mentioned or done

oai_token-act-pair · gpt-3.5-turbo

the term "previously" or its variations in past contexts or references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

eer

-0.71

 Karma

-0.71

alion

-0.69

antics

-0.66

rament

-0.65

 essence

-0.64

 Templ

-0.63

olt

-0.62

onest

-0.62

piration

-0.62

POSITIVE LOGITS

 exting

1.08

cedented

0.85

 unpublished

0.83

 redes

0.83

ãĤ»

0.80

FTWARE

0.80

 compr

0.79

shown

0.78

 authenticated

0.78

 previously

0.78

Activations Density 0.011%

instances where something was previously mentioned or done

the term "previously" or its variations in past contexts or references

No Comments

No Known Activations