INDEX

Explanations

phrases or words mentioning a specific action or event being named or referred to

oai_token-act-pair · gpt-3.5-turbo

instances of the word "called" used 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 @ 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

bilt

-0.79

 istg

-0.71

edia

-0.68

LM

-0.66

LOD

-0.65

yip

-0.65

Cub

-0.64

TPS

-0.63

flix

-0.62

insula

-0.62

POSITIVE LOGITS

call

0.77

calling

0.74

 calling

0.72

@#&

0.71

igraph

0.69

 upon

0.69

 Call

0.68

 attention

0.67

 forth

0.64

 Calling

0.62

Activations Density 0.042%

phrases or words mentioning a specific action or event being named or referred to

instances of the word "called" used in various contexts

No Comments

No Known Activations