INDEX

Explanations

references to serious accusations or charges

oai_token-act-pair · gpt-3.5-turbo

mentions of allegations

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

keys

-0.73

 busiest

-0.70

oppy

-0.70

ARCH

-0.70

patch

-0.69

tz

-0.69

skill

-0.66

arton

-0.65

å¸

-0.64

kin

-0.63

POSITIVE LOGITS

 allegations

1.10

 accusations

0.95

 alleging

0.95

 allegation

0.93

 alleges

0.91

 accusation

0.89

 levied

0.87

 leveled

0.85

 accusing

0.83

 leve

0.77

Activations Density 0.037%

references to serious accusations or charges

mentions of allegations

No Comments

No Known Activations