INDEX

Explanations

references to the name "Masaki" and its variants in different contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ï¸ı

-0.74

sburgh

-0.72

ski

-0.66

BMC

-0.64

orer

-0.64

 Saskatchewan

-0.64

 Ames

-0.63

ship

-0.62

ACTED

-0.61

unes

-0.61

POSITIVE LOGITS

querade

1.26

cul

1.18

quer

1.04

dar

1.04

sey

1.03

ques

1.00

que

0.96

ih

0.94

ala

0.92

qu

0.91

Activations Density 0.004%

references to the name "Masaki" and its variants in different contexts

No Comments

No Known Activations