INDEX

Explanations

phrases related to giving information or sharing quotes, often marked with quotation marks

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols present in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 minim

-0.77

 maxim

-0.73

 decomp

-0.69

 coffin

-0.66

 Voyager

-0.65

 scatter

-0.65

 coast

-0.63

 spinning

-0.63

gib

-0.63

 Curve

-0.62

POSITIVE LOGITS

£

0.91

said

0.82

_>

0.81

especially

0.80

¹

0.79

Â¯

0.79

according

0.76

sure

0.75

drivers

0.75

º

0.74

Activations Density 0.181%

phrases related to giving information or sharing quotes, often marked with quotation marks

special characters or symbols present in the text

No Comments

No Known Activations