INDEX

Explanations

quotations marks followed by emotional or assertive statements

oai_token-act-pair · gpt-3.5-turbo

quotations or dialogue in the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etheless

-0.92

eers

-0.78

erity

-0.75

pec

-0.74

inarily

-0.72

adem

-0.71

Ĥª

-0.70

atical

-0.68

itionally

-0.68

redit

-0.68

POSITIVE LOGITS

said

1.00

 said

0.99

 joked

0.97

 reads

0.96

 says

0.92

 exclaimed

0.89

 tweeted

0.89

 replied

0.89

 recalls

0.85

 remembers

0.85

Activations Density 0.095%

quotations marks followed by emotional or assertive statements

quotations or dialogue in the text

No Comments

No Known Activations