INDEX

Explanations

words containing the substring "use"

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "use."

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

inas

-0.74

inosaur

-0.71

 behold

-0.69

 teasp

-0.65

 whisk

-0.64

ihar

-0.62

etermined

-0.62

ccording

-0.62

 damned

-0.60

ilater

-0.59

POSITIVE LOGITS

ibility

0.85

ppe

0.79

fully

0.76

iculture

0.74

xton

0.71

iveness

0.69

fulness

0.68

FUL

0.68

llo

0.68

ibilities

0.67

Activations Density 0.026%

words containing the substring "use"

references to the concept of "use."

No Comments

No Known Activations