INDEX

Explanations

words related to fatal events or locations

oai_token-act-pair · gpt-3.5-turbo

terms related to fatal incidents or accidents

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

Cloud

-0.74

EY

-0.70

PUT

-0.65

ï¸ı

-0.65

Coll

-0.64

Vert

-0.63

liness

-0.63

ï¸

-0.62

Favorite

-0.62

LD

-0.61

POSITIVE LOGITS

atal

1.39

etheless

0.88

ascus

0.86

idad

0.86

uxe

0.85

anguage

0.84

warn

0.84

ashtra

0.82

ogical

0.80

ania

0.80

Activations Density 0.003%

words related to fatal events or locations

terms related to fatal incidents or accidents

No Comments

No Known Activations