INDEX

Explanations

specific phrases denoting references to files or photographs

oai_token-act-pair · gpt-3.5-turbo

instances of file references or mentions

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

powers

-0.82

aic

-0.81

atform

-0.75

psey

-0.74

anson

-0.73

ilton

-0.73

itational

-0.72

orically

-0.72

ische

-0.71

hhhh

-0.71

POSITIVE LOGITS

 FILE

0.99

 File

0.97

FILE

0.96

Upload

0.88

Path

0.85

upload

0.79

 Photographer

0.78

File

0.76

VIEW

0.75

 PHOTO

0.74

Activations Density 0.016%

specific phrases denoting references to files or photographs

instances of file references or mentions

No Comments

No Known Activations