INDEX

Explanations

phrases indicating membership or inclusion

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "part" in various contexts

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

incinn

-0.76

anche

-0.67

apons

-0.64

asu

-0.62

 shorts

-0.62

 favourite

-0.62

ãĤī

-0.61

 grandchildren

-0.61

 arms

-0.60

stasy

-0.60

POSITIVE LOGITS

 part

0.83

ICLE

0.81

icularly

0.77

heartedly

0.71

ners

0.70

part

0.69

 Reck

0.68

aldehyde

0.68

icles

0.68

 Payton

0.67

Activations Density 0.046%

phrases indicating membership or inclusion

occurrences of the word "part" in various contexts

No Comments

No Known Activations