INDEX

Explanations

numerical values and statistics mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

positive feedback and support in discussions or comments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

thood

-0.77

phans

-0.68

soDeliveryDate

-0.64

 ageing

-0.64

iferation

-0.63

 starve

-0.62

allows

-0.62

brance

-0.61

bis

-0.61

¥µ

-0.61

POSITIVE LOGITS

 unanimously

0.93

 ranging

0.90

 apiece

0.87

 indicating

0.79

ranging

0.78

 succinct

0.77

 varied

0.77

 includ

0.77

0.76

 respectively

0.76

Activations Density 0.685%

numerical values and statistics mentioned in the text

positive feedback and support in discussions or comments

No Comments

No Known Activations