INDEX

Explanations

references to a specific location, namely "Auburn"

oai_token-act-pair · gpt-3.5-turbo

mentions of "Auburn" and related entities or individuals 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

arded

-0.77

iculty

-0.75

rooms

-0.64

acqu

-0.63

Batman

-0.63

 Madden

-0.63

ribut

-0.63

DCS

-0.63

 Bray

-0.63

igger

-0.63

POSITIVE LOGITS

 unden

0.80

oir

0.78

ous

0.73

comings

0.72

cia

0.72

zee

0.71

thur

0.71

sts

0.70

thood

0.70

ously

0.70

Activations Density 0.045%

references to a specific location, namely "Auburn"

mentions of "Auburn" and related entities or individuals in various contexts

No Comments

No Known Activations