INDEX

Explanations

phrases in a specific language or encoding pattern

oai_token-act-pair · gpt-3.5-turbo

special characters or unusual symbols

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

best

-0.74

drawn

-0.74

geries

-0.73

gotten

-0.70

 similarity

-0.69

 simplest

-0.68

itars

-0.68

 best

-0.67

lycer

-0.65

pert

-0.65

POSITIVE LOGITS

į

1.61

ÃįÃį

1.03

ķ

1.01

à¤

0.98

£

0.98

Į

0.95

Ģ

0.93

°

0.92

ÑĤ

0.92

Ĭ

0.92

Activations Density 0.007%

phrases in a specific language or encoding pattern

special characters or unusual symbols

No Comments

No Known Activations