INDEX

Explanations

names or words ending in "lin"

oai_token-act-pair · gpt-3.5-turbo

the occurrences of the word "lin"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Domain

-0.71

 takedown

-0.66

 domain

-0.65

 handheld

-0.65

 FISA

-0.63

 seizures

-0.63

ETF

-0.62

AK

-0.62

 custody

-0.61

 impunity

-0.60

POSITIVE LOGITS

lin

4.49

lins

2.64

LIN

2.12

liner

1.55

lan

1.53

len

1.52

lyn

1.46

ling

1.43

line

1.37

lon

1.32

Activations Density 0.023%

names or words ending in "lin"

the occurrences of the word "lin"

No Comments

No Known Activations