INDEX

Explanations

proper nouns related to a specific location, possibly indicating a focus on news or events happening in that location

oai_token-act-pair · gpt-3.5-turbo

occurrences of the token "rin."

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

 Seym

-0.82

eatures

-0.76

gom

-0.73

uries

-0.66

oine

-0.65

arcer

-0.63

 Hurricanes

-0.63

inctions

-0.63

gettable

-0.62

 therap

-0.62

POSITIVE LOGITS

ned

1.00

ners

0.92

ergic

0.91

åĤ

0.87

icum

0.83

othing

0.83

ja

0.83

lich

0.82

voy

0.80

hof

0.80

Activations Density 0.012%

proper nouns related to a specific location, possibly indicating a focus on news or events happening in that location

occurrences of the token "rin."

No Comments

No Known Activations