INDEX

Explanations

mentions of executive leadership roles

oai_token-act-pair · gpt-3.5-turbo

titles of individuals in executive positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

etimes

-0.75

apo

-0.71

aepernick

-0.71

ONES

-0.70

elight

-0.70

nikov

-0.69

 Pastebin

-0.66

 Frenzy

-0.66

pees

-0.66

ÅŁ

-0.65

POSITIVE LOGITS

 executive

1.29

 engineer

1.13

 negotiator

1.11

 architect

1.09

 strategist

1.03

 advisor

0.99

 investigator

0.98

 economist

0.96

 adviser

0.95

 scientist

0.94

Activations Density 0.034%

mentions of executive leadership roles

titles of individuals in executive positions

No Comments

No Known Activations