INDEX

Explanations

terms related to power consumption and energy output

oai_token-act-pair · gpt-3.5-turbo

mentions of measurements or quantities in a technical or sports context

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

 chambers

-0.74

 partial

-0.73

 borderline

-0.68

 chamber

-0.66

 wired

-0.66

 deepening

-0.66

 detailed

-0.65

lur

-0.65

 rabbit

-0.65

rises

-0.65

POSITIVE LOGITS

å£«

1.00

theless

0.92

ograph

0.91

andise

0.89

lihood

0.86

iscopal

0.85

obile

0.83

Leaks

0.82

iversary

0.81

Mania

0.81

Activations Density 0.017%

terms related to power consumption and energy output

mentions of measurements or quantities in a technical or sports context

No Comments

No Known Activations