INDEX

Explanations

phrases with actions or activities mentioned in a conversational context

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, particularly quotes and apostrophes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 polyg

-0.68

peg

-0.67

 lapse

-0.66

 showers

-0.63

 detail

-0.62

 sympathetic

-0.62

ensical

-0.61

 elev

-0.61

 linger

-0.60

 Objects

-0.60

POSITIVE LOGITS

til

1.49

Cause

1.44

cause

1.32

Mech

1.13

tis

1.12

Interstitial

1.08

Course

1.07

neath

1.00

round

0.98

Em

0.97

Activations Density 0.038%

phrases with actions or activities mentioned in a conversational context

punctuation marks, particularly quotes and apostrophes

No Comments

No Known Activations