INDEX

Explanations

phrases indicating a choice or possibility

oai_token-act-pair · gpt-3.5-turbo

phrases indicating conditionality and potentiality

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

ratulations

-0.70

itiz

-0.70

 congratulations

-0.63

athing

-0.61

ãĥĥãĥĪ

-0.59

orer

-0.59

phabet

-0.58

eur

-0.58

 Gone

-0.56

itary

-0.56

POSITIVE LOGITS

 expense

0.99

 moment

0.97

 glance

0.90

 rate

0.86

mom

0.83

 behest

0.82

 point

0.82

 cost

0.82

point

0.81

cost

0.80

Activations Density 0.033%

phrases indicating a choice or possibility

phrases indicating conditionality and potentiality

No Comments

No Known Activations