INDEX

Explanations

words related to making tasks easier or more manageable

oai_token-act-pair · gpt-3.5-turbo

expressions indicating that something is made simpler or more efficient

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

eters

-0.72

inals

-0.69

Origin

-0.69

Saint

-0.67

ARE

-0.66

 Anthem

-0.66

borg

-0.64

hips

-0.63

ulations

-0.63

ember

-0.62

POSITIVE LOGITS

 than

1.06

 forgiving

0.89

 prey

0.84

 "$:/

0.79

 compr

0.77

 manageable

0.77

iquid

0.72

 readable

0.71

 digestion

0.68

housing

0.68

Activations Density 0.015%

words related to making tasks easier or more manageable

expressions indicating that something is made simpler or more efficient

No Comments

No Known Activations