INDEX

Explanations

phrases related to being subjected to some form of action or treatment

oai_token-act-pair · gpt-3.5-turbo

instances of being subjected to actions or situations against one's will

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Forge

-0.65

 genesis

-0.62

 Alps

-0.61

ENCE

-0.61

 predecessors

-0.60

 Stability

-0.60

 dedication

-0.60

 ingenuity

-0.60

 Beaver

-0.60

 convergence

-0.59

POSITIVE LOGITS

 aback

1.02

 captive

0.95

 hostage

0.94

 prisoner

0.88

 deported

0.87

otyp

0.86

cffffcc

0.80

quished

0.79

 Reincarnated

0.79

leep

0.78

Activations Density 0.279%

phrases related to being subjected to some form of action or treatment

instances of being subjected to actions or situations against one's will

No Comments

No Known Activations