INDEX

Explanations

references to previous statements or mentions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating quotations or references to statements made earlier

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

 Lauder

-0.90

âĺħ

-0.64

airs

-0.64

 imprint

-0.63

blem

-0.63

=~=~

-0.62

Mouse

-0.61

icons

-0.60

arel

-0.60

imi

-0.59

POSITIVE LOGITS

 earlier

0.74

 goodbye

0.73

ratulations

0.72

EStream

0.69

bye

0.63

escription

0.60

 Izan

0.60

 theat

0.59

 categ

0.58

 caveats

0.58

Activations Density 0.101%

references to previous statements or mentions

phrases indicating quotations or references to statements made earlier

No Comments

No Known Activations