INDEX

Explanations

words related to accusations and critiques of actions by individuals in authority positions

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

MpServer

-0.78

ãĥīãĥ©ãĤ´ãĥ³

-0.74

CHO

-0.66

Magikarp

-0.65

senal

-0.64

ADRA

-0.61

Issue

-0.61

WHERE

-0.59

xon

-0.59

AAF

-0.59

POSITIVE LOGITS

shire

1.13

ords

1.12

enegger

1.10

erent

1.04

ees

1.00

irms

0.98

ington

0.97

rey

0.95

yre

0.94

sburgh

0.92

Activations Density 0.009%

words related to accusations and critiques of actions by individuals in authority positions

No Comments

No Known Activations