INDEX

Explanations

phrases related to taking responsibility for actions

oai_token-act-pair · gpt-3.5-turbo

concepts and discussions surrounding the notion of accountability and personal responsibility

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ugen

-0.76

wikipedia

-0.73

arthed

-0.69

tering

-0.69

glers

-0.67

 Racer

-0.67

inton

-0.67

ropolis

-0.66

corn

-0.65

tein

-0.65

POSITIVE LOGITS

 responsibilities

0.89

 incurred

0.86

 responsibility

0.84

 owed

0.80

lessness

0.79

 delegated

0.79

 forg

0.79

 accountable

0.77

 Responsibility

0.76

 entrusted

0.75

Activations Density 0.037%

phrases related to taking responsibility for actions

concepts and discussions surrounding the notion of accountability and personal responsibility

No Comments

No Known Activations