INDEX

Explanations

instances of identical or nearly identical things in different contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "identical" and related concepts of sameness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

raq

-0.78

stra

-0.78

erest

-0.75

âĵĺ

-0.71

sta

-0.71

================================================================

-0.70

HI

-0.70

bara

-0.69

 rhet

-0.68

uay

-0.68

POSITIVE LOGITS

 twins

1.20

 twin

0.90

icut

0.84

 identical

0.80

lihood

0.78

 sized

0.74

 pairs

0.72

etrical

0.70

 digits

0.70

 copies

0.69

Activations Density 0.031%

instances of identical or nearly identical things in different contexts

instances of the word "identical" and related concepts of sameness

No Comments

No Known Activations