INDEX

Explanations

names of people, places, and organizations

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

|--

-0.61

 veter

-0.56

 ILCS

-0.55

¿½

-0.54

pires

-0.54

Ĥ¬

-0.52

(%)

-0.52

 Undead

-0.51

"}],"

-0.51

 Emin

-0.50

POSITIVE LOGITS

 trolls

0.56

 reconsider

0.55

 bandwagon

0.54

opp

0.54

 skeptics

0.53

 privatization

0.53

panel

0.50

 boycot

0.50

chedel

0.49

 skept

0.48

Activations Density 0.942%

names of people, places, and organizations

proper nouns, particularly names and organizations

No Comments

No Known Activations