INDEX

Explanations

words related to extreme negative emotions such as hatred

oai_token-act-pair · gpt-3.5-turbo

references to hatred and its various manifestations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

UNCH

-0.82

ODE

-0.81

 helicop

-0.75

umm

-0.69

 stabilize

-0.66

debian

-0.65

å¸

-0.64

change

-0.64

aqu

-0.63

Interstitial

-0.63

POSITIVE LOGITS

 hatred

1.01

 prejudice

0.94

 towards

0.88

 vengeance

0.84

ãĥĨ

0.81

 toward

0.80

fully

0.79

lessly

0.76

 spew

0.74

 revenge

0.72

Activations Density 0.020%

words related to extreme negative emotions such as hatred

references to hatred and its various manifestations

No Comments

No Known Activations