INDEX

Explanations

phrases related to connections and relationships between people or entities

oai_token-act-pair · gpt-3.5-turbo

phrases that contain slashes, indicating divisions or categories in text

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

 swell

-0.80

 square

-0.78

 lifetime

-0.78

 densely

-0.77

 sincerely

-0.75

 Chao

-0.74

 younger

-0.74

 mate

-0.73

 leaflets

-0.73

 breed

-0.73

POSITIVE LOGITS

whatever

1.67

etc

1.67

dist

1.46

coll

1.46

trans

1.45

tem

1.44

super

1.42

non

1.40

control

1.40

workshop

1.40

Activations Density 0.052%

phrases related to connections and relationships between people or entities

phrases that contain slashes, indicating divisions or categories in text

No Comments

No Known Activations