INDEX

Explanations

words related to disapproval or criticism

oai_token-act-pair · gpt-3.5-turbo

words related to public relations or promotional content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Sacrament

-0.73

 Winds

-0.69

BILITY

-0.67

ORED

-0.63

 persistence

-0.63

 Eagle

-0.63

 Royale

-0.62

born

-0.62

 Americas

-0.61

croft

-0.60

POSITIVE LOGITS

imate

1.12

uning

1.10

imes

1.08

arians

1.06

ices

0.97

atical

0.96

icy

0.94

ams

0.94

icks

0.92

asion

0.92

Activations Density 0.011%

words related to disapproval or criticism

words related to public relations or promotional content

No Comments

No Known Activations