INDEX

Explanations

phrases related to candidates or competitors for a specific position or job

oai_token-act-pair · gpt-3.5-turbo

phrases indicating competition for positions or roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

raq

-1.02

soever

-0.76

ighed

-0.74

entimes

-0.74

arrow

-0.73

uristic

-0.72

heit

-0.72

Î³

-0.71

speak

-0.69

ient

-0.69

POSITIVE LOGITS

 nomination

1.11

 supremacy

1.04

 Best

0.94

 presidency

0.89

 throne

0.88

 nominations

0.88

 immortality

0.84

 contention

0.83

 best

0.82

 leadership

0.82

Activations Density 0.122%

phrases related to candidates or competitors for a specific position or job

phrases indicating competition for positions or roles

No Comments

No Known Activations