INDEX

Explanations

mentions of specific symbols or characters

oai_token-act-pair · gpt-3.5-turbo

instances of the word "didn't" or its variations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 warp

-0.67

 pyramid

-0.67

 chained

-0.66

 Lancaster

-0.66

 Rampage

-0.65

 Hats

-0.65

 Pony

-0.63

 draped

-0.62

 Scarlet

-0.62

Yor

-0.61

POSITIVE LOGITS

ï¸ı

1.08

vernment

1.05

ulty

1.03

Â¯Â¯

1.00

ufact

0.99

conom

0.98

Â£

0.97

efe

0.95

ember

0.90

Â¢

0.88

Activations Density 0.353%

mentions of specific symbols or characters

instances of the word "didn't" or its variations

No Comments

No Known Activations