INDEX

Explanations

terms related to law enforcement and drugs

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names associated with individuals and organizations

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

oning

-0.79

oned

-0.73

OTS

-0.71

rosse

-0.67

fare

-0.66

 Templ

-0.65

âĶĢâĶĢâĶĢâĶĢ

-0.65

aign

-0.65

UID

-0.64

offs

-0.64

POSITIVE LOGITS

uggets

0.91

iferation

0.91

emonic

0.89

ucle

0.86

swer

0.86

guyen

0.85

issan

0.84

ominated

0.80

jad

0.79

uchin

0.78

Activations Density 0.071%

terms related to law enforcement and drugs

proper nouns and names associated with individuals and organizations

No Comments

No Known Activations