INDEX

Explanations

phrases related to describing or detailing specific aspects or elements

oai_token-act-pair · gpt-3.5-turbo

constructions that indicate the extent or manner of actions or influences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Advertisement

-0.68

©¶æ¥µ

-0.56

 compat

-0.55

 Mile

-0.55

Cly

-0.52

ŃĶ

-0.52

 TOTAL

-0.51

olulu

-0.50

OSH

-0.50

ħĭ

-0.49

POSITIVE LOGITS

 which

1.20

which

1.08

 whence

0.91

 Which

0.87

rium

0.81

 wherein

0.80

 guiActiveUnfocused

0.77

itability

0.73

Which

0.72

 whom

0.70

Activations Density 0.407%

phrases related to describing or detailing specific aspects or elements

constructions that indicate the extent or manner of actions or influences

No Comments

No Known Activations