INDEX

Explanations

concepts related to power dynamics and manipulation through language

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ancies

-0.82

omes

-0.77

 Scenes

-0.76

 Indigo

-0.76

 Clover

-0.72

sbm

-0.71

tis

-0.71

 batches

-0.70

tones

-0.69

OTOS

-0.68

POSITIVE LOGITS

 circumvent

0.96

 achieving

0.90

 deterrence

0.89

 coping

0.88

 dissemin

0.84

 recourse

0.83

 combating

0.81

 measure

0.80

 aiming

0.78

 lever

0.78

Activations Density 0.159%

concepts related to power dynamics and manipulation through language

No Comments

No Known Activations