INDEX

Explanations

keywords related to markers or indicators

oai_token-act-pair · gpt-3.5-turbo

terms related to markers used for classification or identification

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

orld

-0.84

erest

-0.83

irl

-0.81

ews

-0.77

obbies

-0.76

ategory

-0.74

awar

-0.72

alez

-0.71

sid

-0.71

oÄŁ

-0.70

POSITIVE LOGITS

 marker

1.56

 markers

1.39

 plaque

0.88

posts

0.87

 marking

0.85

holder

0.81

 indicating

0.74

 Twain

0.72

 dotted

0.72

 flare

0.71

Activations Density 0.005%

keywords related to markers or indicators

terms related to markers used for classification or identification

No Comments

No Known Activations