INDEX

Explanations

non-human language characters or symbols

oai_token-act-pair · gpt-3.5-turbo

special characters and symbols often associated with formatting or encoding issues

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

 recogn

-0.69

theless

-0.68

Reloaded

-0.66

 diagn

-0.65

Mol

-0.64

 today

-0.61

 comprom

-0.60

 Tonight

-0.59

 hypnot

-0.59

 Tyrann

-0.59

POSITIVE LOGITS

ĵ

1.96

Ķ

1.87

¦

1.84

Ĳ

1.84

ĥ

1.77

ª

1.76

ĺ

1.75

Ĵ

1.74

³

1.74

ĳ

1.74

Activations Density 0.037%

non-human language characters or symbols

special characters and symbols often associated with formatting or encoding issues

No Comments

No Known Activations