INDEX

Explanations

the word "bench" within sentences

oai_token-act-pair · gpt-3.5-turbo

references to "bench" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Operation

-0.82

 Capitalism

-0.75

 Carnage

-0.70

Ca

-0.69

 Spir

-0.68

 Typh

-0.67

Credit

-0.64

 escaped

-0.62

opsis

-0.62

 Mond

-0.62

POSITIVE LOGITS

 bench

4.13

 benches

2.94

 Bench

2.72

bench

2.31

Bench

2.25

ben

1.21

 sidelines

1.14

 bullpen

1.13

 podium

1.13

 sideline

1.09

Activations Density 0.009%

the word "bench" within sentences

references to "bench" in various contexts

No Comments

No Known Activations