INDEX

Explanations

quantitative measures related to specific areas or spaces

oai_token-act-pair · gpt-3.5-turbo

measurements of area, specifically in square feet and square meters

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

 bearer

-0.83

 GOODMAN

-0.83

 medi

-0.70

 responsible

-0.68

 Glory

-0.66

OLOG

-0.66

 Enforcement

-0.62

 directing

-0.62

IBLE

-0.61

HOME

-0.60

POSITIVE LOGITS

sq

1.03

sq

1.02

querque

0.79

Äį

0.78

inct

0.78

Ã»

0.74

vette

0.74

ropy

0.74

avin

0.73

rt

0.73

Activations Density 0.004%

quantitative measures related to specific areas or spaces

measurements of area, specifically in square feet and square meters

No Comments

No Known Activations