INDEX

Explanations

parentheses indicating statistics or numerical information

oai_token-act-pair · gpt-3.5-turbo

instances of parentheses used in stats or scores

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Ceres

-0.79

nic

-0.78

bro

-0.78

 wildlife

-0.76

ram

-0.76

 barr

-0.73

orn

-0.72

 breed

-0.72

 swell

-0.72

 settlement

-0.72

POSITIVE LOGITS

six

1.51

seven

1.48

three

1.46

fourth

1.45

four

1.45

albeit

1.45

both

1.45

second

1.45

nine

1.44

minimum

1.44

Activations Density 0.092%

parentheses indicating statistics or numerical information

instances of parentheses used in stats or scores

No Comments

No Known Activations