INDEX

Explanations

personal pronouns followed by positive statements about one's actions or beliefs

oai_token-act-pair · gpt-3.5-turbo

pronouns and references to individuals, particularly in contexts of achievement and recognition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 beware

-0.71

 longing

-0.65

 Dragonbound

-0.61

RET

-0.60

quartered

-0.60

 approximation

-0.59

URI

-0.59

ranged

-0.59

ulsion

-0.58

onde

-0.56

POSITIVE LOGITS

 accomplished

1.03

did

0.93

did

0.89

 accomplish

0.86

 done

0.85

've

0.83

 wrought

0.81

do

0.79

 achieved

0.77

done

0.76

Activations Density 0.110%

personal pronouns followed by positive statements about one's actions or beliefs

pronouns and references to individuals, particularly in contexts of achievement and recognition

No Comments

No Known Activations