INDEX

Explanations

words related to names ending in 'der'

oai_token-act-pair · gpt-3.5-turbo

a recurring pattern involving the word "der" in various contexts

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

ainted

-0.62

Pros

-0.57

urally

-0.57

heres

-0.57

oufl

-0.56

INTER

-0.56

eno

-0.54

 Reconstruction

-0.54

 African

-0.54

EH

-0.53

POSITIVE LOGITS

dash

0.97

rors

0.89

cair

0.87

minster

0.84

der

0.84

izont

0.83

netic

0.82

theless

0.82

ocket

0.81

iving

0.79

Activations Density 0.009%

words related to names ending in 'der'

a recurring pattern involving the word "der" in various contexts

No Comments

No Known Activations