INDEX

Explanations

numerical values or hashtags

oai_token-act-pair · gpt-3.5-turbo

symbols and numeric data, particularly related to statistics or formatting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Lerner

-0.77

enne

-0.75

-0.71

-0.70

 Yankee

-0.70

ese

-0.68

-0.67

-0.65

-0.64

POSITIVE LOGITS

0.99

0.93

0.79

Saturday

0.76

0.74

0.73

0.72

execute

0.72

0.71

Activations Density 0.068%

numerical values or hashtags

symbols and numeric data, particularly related to statistics or formatting

No Comments

No Known Activations