INDEX

Explanations

phrases related to hidden or underlying issues

oai_token-act-pair · gpt-3.5-turbo

phrases indicating locations or positions within contexts

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

nces

-0.96

xual

-0.84

unctions

-0.78

uthor

-0.77

cientious

-0.77

itives

-0.76

iors

-0.75

idents

-0.74

members

-0.74

AUD

-0.73

POSITIVE LOGITS

 coffin

1.03

 proverbial

1.02

 iceberg

0.89

 wedge

0.85

hay

0.78

 rotten

0.78

 bucket

0.76

 puzzle

0.75

 apple

0.74

pie

0.74

Activations Density 0.344%

phrases related to hidden or underlying issues

phrases indicating locations or positions within contexts

No Comments

No Known Activations