INDEX

Explanations

adjectives ending in 'ly', primarily focusing on emotional or physical attributes

oai_token-act-pair · gpt-3.5-turbo

adjectives ending in "ly."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

arten

-0.81

uther

-0.76

Specific

-0.72

involved

-0.70

ICAN

-0.69

liga

-0.69

ĨĴ

-0.69

ague

-0.67

anta

-0.66

ighton

-0.66

POSITIVE LOGITS

old

0.85

 grin

0.84

ol

0.80

ly

0.80

 humour

0.80

hearted

0.79

 glances

0.78

ishly

0.78

 eyed

0.76

 minded

0.75

Activations Density 0.085%

adjectives ending in 'ly', primarily focusing on emotional or physical attributes

adjectives ending in "ly."

No Comments

No Known Activations