INDEX

Explanations

patterns or sequences of characters, likely related to programming code or technical information

oai_token-act-pair · gpt-3.5-turbo

numeric data and formatting symbols

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

 Frankfurt

-0.79

 Finn

-0.76

 Franch

-0.76

 PASS

-0.75

 arri

-0.73

CRA

-0.72

Bri

-0.72

Hab

-0.71

 Sinn

-0.70

HAL

-0.69

POSITIVE LOGITS

2.35

1.73

xs

1.69

xc

1.65

xf

1.59

xes

1.57

xff

1.55

xa

1.54

xd

1.54

xe

1.54

Activations Density 0.077%

patterns or sequences of characters, likely related to programming code or technical information

numeric data and formatting symbols

No Comments

No Known Activations