INDEX

Explanations

phrases related to programming or technical instructions

oai_token-act-pair · gpt-3.5-turbo

references to performance and entertainment elements

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ward

-1.09

 Ward

-0.88

 Levi

-0.87

Sab

-0.84

 Stra

-0.83

 ward

-0.81

 Shib

-0.77

 Daniels

-0.77

 Ross

-0.76

Flo

-0.75

POSITIVE LOGITS

 Classic

1.25

classic

1.25

Classic

1.16

 Classics

1.00

 classic

0.97

 Chapman

0.96

 Bour

0.90

historic

0.90

sonian

0.88

traditional

0.87

Activations Density 0.344%

phrases related to programming or technical instructions

references to performance and entertainment elements

No Comments

No Known Activations