INDEX

Explanations

negative portrayals and characterizations of specific political groups or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ¡

-0.70

ãĤ©

-0.69

uyomi

-0.64

vironment

-0.60

UTF

-0.60

QUI

-0.59

STEM

-0.57

osterone

-0.57

CAST

-0.55

 Alloy

-0.55

POSITIVE LOGITS

akening

0.66

imester

0.61

tsy

0.61

vu

0.60

 Clyde

0.59

 necks

0.59

atari

0.59

reviewed

0.58

 Nights

0.56

 limb

0.56

Activations Density 0.044%

negative portrayals and characterizations of specific political groups or individuals

No Comments

No Known Activations