INDEX

Explanations

phrases related to acts of aggression

oai_token-act-pair · gpt-3.5-turbo

instances of the word "aggression" and related forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

FORMATION

-0.82

lev

-0.69

Vert

-0.68

ectar

-0.68

aver

-0.67

zl

-0.67

HCR

-0.67

âĢ¢âĢ¢

-0.67

orah

-0.66

Rated

-0.66

POSITIVE LOGITS

 aggression

1.12

 aggress

0.93

 provocation

0.90

iveness

0.86

 escalation

0.82

 against

0.80

 towards

0.76

 posture

0.75

 toward

0.74

 provoked

0.73

Activations Density 0.019%

phrases related to acts of aggression

instances of the word "aggression" and related forms

No Comments

No Known Activations