INDEX

Explanations

words related to criticism or caution

oai_token-act-pair · gpt-3.5-turbo

references to man-made phenomena or entities

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

Ces

-0.69

 Chambers

-0.66

 Cotton

-0.66

 Grimes

-0.66

 bitters

-0.64

 Ambrose

-0.63

 waning

-0.62

 Pell

-0.61

 Euph

-0.61

 remaining

-0.61

POSITIVE LOGITS

centered

1.43

centric

1.30

shaped

1.30

sized

1.24

friendly

1.22

made

1.20

readable

1.20

oriented

1.17

specific

1.17

machine

1.14

Activations Density 0.067%

words related to criticism or caution

references to man-made phenomena or entities

No Comments

No Known Activations