INDEX

Explanations

mentions of stalking behavior

oai_token-act-pair · gpt-3.5-turbo

terms related to stalking and creepy behavior

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 immersion

-0.74

ption

-0.72

gae

-0.68

ilts

-0.68

andum

-0.68

itars

-0.67

ptive

-0.67

iyah

-0.66

ederation

-0.65

 Rite

-0.65

POSITIVE LOGITS

 Creep

1.09

 stalking

1.01

 stalk

0.96

ingly

0.91

crow

0.90

ritch

0.81

lords

0.79

 creep

0.79

 Detect

0.77

 harass

0.75

Activations Density 0.037%

mentions of stalking behavior

terms related to stalking and creepy behavior

No Comments

No Known Activations