INDEX

Explanations

words related to sparking intense reactions, debates, and movements

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the initiation of discussions or reactions to significant events or issues

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

 Anat

-0.76

 undone

-0.70

ACTED

-0.69

 Definitive

-0.67

 potatoes

-0.67

 disapp

-0.65

 implements

-0.64

 peeled

-0.64

 flats

-0.64

 repealed

-0.62

POSITIVE LOGITS

inki

0.88

ire

0.79

riott

0.73

 disbelief

0.72

 wrath

0.72

Demand

0.71

aign

0.70

 downfall

0.69

 frenzy

0.69

ependence

0.68

Activations Density 0.283%

words related to sparking intense reactions, debates, and movements

phrases that indicate the initiation of discussions or reactions to significant events or issues

No Comments

No Known Activations