INDEX

Explanations

the word "red" in various contexts

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_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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ernel

-0.79

UGH

-0.77

 Lank

-0.70

XT

-0.70

awaru

-0.70

Ö¼

-0.70

Reloaded

-0.69

ILA

-0.69

agall

-0.66

 incorpor

-0.64

POSITIVE LOGITS

neck

1.22

efined

1.21

oubt

1.16

iscovered

1.13

oub

1.12

irection

1.10

rawn

1.08

ucing

1.08

iscover

1.06

uced

1.05

Activations Density 0.029%

the word "red" in various contexts

No Comments

No Known Activations