INDEX

Explanations

letters and symbols occurring in patterns

oai_token-act-pair · gpt-3.5-turbo

instances of unique or unexpected punctuation and special characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

hement

-0.72

isconsin

-0.63

abwe

-0.63

akings

-0.62

 footing

-0.61

anes

-0.59

acan

-0.58

emark

-0.57

 elimination

-0.57

isons

-0.56

POSITIVE LOGITS

--------------------------------------------------------

0.92

Reward

0.89

à¼

0.89

Synopsis

0.88

Plug

0.85

Relations

0.82

âĶĢâĶĢâĶĢâĶĢ

0.82

ãģı

0.81

ÙĦ

0.80

 Norn

0.79

Activations Density 0.269%

letters and symbols occurring in patterns

instances of unique or unexpected punctuation and special characters

No Comments

No Known Activations