INDEX

Explanations

phrases indicating intention or purpose

oai_token-act-pair · gpt-3.5-turbo

phrases indicating intentions or actions related to personal and collective experiences

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

acies

-0.74

ventions

-0.70

sequ

-0.69

zie

-0.69

artisan

-0.66

Cas

-0.66

orio

-0.65

ilings

-0.65

ournal

-0.65

cffff

-0.63

POSITIVE LOGITS

liest

0.89

hest

0.71

 longest

0.68

 proverb

0.63

chin

0.62

iest

0.62

EMENT

0.59

 richest

0.59

 "$:/

0.58

 behav

0.58

Activations Density 0.324%

phrases indicating intention or purpose

phrases indicating intentions or actions related to personal and collective experiences

No Comments

No Known Activations