INDEX

Explanations

phrases related to forceful actions or coercion

oai_token-act-pair · gpt-3.5-turbo

phrases indicating coercion or being forced into actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

/-

-0.73

////////////////////////////////

-0.71

icious

-0.66

fortunately

-0.64

vironment

-0.63

İ

-0.61

ergy

-0.61

volent

-0.60

laughter

-0.60

 Prediction

-0.60

POSITIVE LOGITS

 confront

1.22

 endure

1.21

 undergo

1.21

 rethink

1.09

 obey

1.06

 undertake

1.05

 abandon

1.04

 relocate

1.03

 conform

1.02

 rely

1.02

Activations Density 0.108%

phrases related to forceful actions or coercion

phrases indicating coercion or being forced into actions

No Comments

No Known Activations