INDEX

Explanations

terms related to incremental changes or improvements

oai_token-act-pair · gpt-3.5-turbo

terms related to incremental changes or improvements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Aval

-0.80

rigan

-0.73

buster

-0.72

argon

-0.70

wagen

-0.64

opsis

-0.62

 Deborah

-0.62

NetMessage

-0.61

ridge

-0.61

 Hawaiian

-0.61

POSITIVE LOGITS

mental

1.22

ments

1.06

ment

0.98

asing

0.98

 incre

0.98

 increment

0.92

mented

0.89

 Incre

0.85

mble

0.83

ally

0.81

Activations Density 0.020%

terms related to incremental changes or improvements

terms related to incremental changes or improvements

No Comments

No Known Activations