INDEX

Explanations

names of specific individuals

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

actionDate

-0.65

 Stella

-0.58

ãĥ¼ãĥ³

-0.57

Lex

-0.55

 MDMA

-0.55

 EDITION

-0.54

ORT

-0.54

ĭ

-0.54

WCS

-0.54

CAP

-0.53

POSITIVE LOGITS

 himself

1.23

's

1.16

 testified

1.07

enegger

1.07

 admits

1.06

 wrote

1.01

 remembers

1.00

 insists

0.99

 concedes

0.98

 believes

0.97

Activations Density 1.012%

names of specific individuals

No Comments

No Known Activations