INDEX

Explanations

mentions of corruption

oai_token-act-pair · gpt-3.5-turbo

references to corruption in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

TRY

-0.85

 Flavoring

-0.76

WAYS

-0.74

 Bears

-0.72

ynthesis

-0.71

ttp

-0.69

ãĤ´ãĥ³

-0.69

CAR

-0.68

IGHT

-0.68

cknowled

-0.68

POSITIVE LOGITS

 corrupt

1.28

 corrupted

0.98

 corruption

0.96

 undermin

0.85

ingly

0.79

ible

0.79

 overse

0.77

ulent

0.75

 dece

0.74

 Corruption

0.73

Activations Density 0.008%

mentions of corruption

references to corruption in various contexts

No Comments

No Known Activations