INDEX

Explanations

phrases related to earning achievements or recognition

oai_token-act-pair · gpt-3.5-turbo

instances of reputation and achievement

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

anwhile

-0.89

ESE

-0.66

 forums

-0.65

 processes

-0.64

soDeliveryDate

-0.63

hari

-0.63

ricting

-0.62

eki

-0.62

 Asians

-0.61

 Boards

-0.60

POSITIVE LOGITS

 coveted

0.90

ire

0.88

 notoriety

0.85

 renown

0.77

 livelihood

0.76

 nickname

0.73

 certification

0.72

 Citation

0.72

 foothold

0.69

 firsthand

0.69

Activations Density 0.165%

phrases related to earning achievements or recognition

instances of reputation and achievement

No Comments

No Known Activations