INDEX

Explanations

terms related to government, politics, espionage, and military activities

oai_token-act-pair · gpt-3.5-turbo

proper nouns and specific group identifiers related to entities and demographics

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

eanor

-0.75

ibaba

-0.73

uyomi

-0.69

vable

-0.67

itars

-0.67

chwitz

-0.65

accessible

-0.63

 positives

-0.62

comings

-0.62

ashtra

-0.62

POSITIVE LOGITS

º

0.60

¦

0.59

 contractor

0.59

 Writers

0.59

 Logged

0.57

 guiActiveUn

0.56

 intermediary

0.56

 commentator

0.56

ADA

0.56

eman

0.56

Activations Density 0.386%

terms related to government, politics, espionage, and military activities

proper nouns and specific group identifiers related to entities and demographics

No Comments

No Known Activations