INDEX

Explanations

phrases related to locations or directions

oai_token-act-pair · gpt-3.5-turbo

references to corners, particularly in physical or spatial contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ufact

-0.83

egal

-0.83

orthy

-0.76

inen

-0.76

atoon

-0.75

issance

-0.75

hur

-0.71

ploy

-0.71

wow

-0.71

ignty

-0.71

POSITIVE LOGITS

stones

1.20

bars

0.88

busters

0.86

 corner

0.84

istas

0.79

 corners

0.78

naire

0.74

nered

0.74

ing

0.73

chester

0.73

Activations Density 0.034%

phrases related to locations or directions

references to corners, particularly in physical or spatial contexts

No Comments

No Known Activations