INDEX

Explanations

terms related to the beginning or start of something, like 'onset', 'outset', and 'commencement'

oai_token-act-pair · gpt-3.5-turbo

terms related to beginnings or starting points

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

lean

-0.80

abase

-0.72

este

-0.70

ald

-0.70

atti

-0.70

ged

-0.69

arov

-0.68

rats

-0.67

verbs

-0.67

orde

-0.66

POSITIVE LOGITS

 onset

1.09

 puberty

0.92

 commencement

0.82

 date

0.80

cffffcc

0.76

wcs

0.72

APR

0.71

 Date

0.71

 unlaw

0.71

 commence

0.67

Activations Density 0.011%

terms related to the beginning or start of something, like 'onset', 'outset', and 'commencement'

terms related to beginnings or starting points

No Comments

No Known Activations