INDEX

Explanations

names or words related to people or places

oai_token-act-pair · gpt-3.5-turbo

names and terms related to specific geographic or fictional locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 rooting

-0.69

 stump

-0.68

 provision

-0.64

 unsett

-0.64

 morbid

-0.63

 metic

-0.62

CRC

-0.62

 departing

-0.62

 bailed

-0.60

avez

-0.59

POSITIVE LOGITS

gard

1.27

ner

1.07

gart

1.07

nings

0.96

ners

0.93

geist

0.93

ens

0.90

ener

0.88

ians

0.88

ian

0.87

Activations Density 0.017%

names or words related to people or places

names and terms related to specific geographic or fictional locations

No Comments

No Known Activations