INDEX

Explanations

descriptions of places and people in narratives

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

asio

-0.78

ESE

-0.69

obia

-0.68

izon

-0.67

WHERE

-0.67

YS

-0.66

Players

-0.65

vale

-0.64

ramids

-0.64

,—

-0.64

POSITIVE LOGITS

 albeit

1.22

um

0.89

albeit

0.85

huh

0.85

 charismatic

0.84

uh

0.84

 supra

0.82

 technologically

0.79

 consensual

0.78

 colourful

0.77

Activations Density 0.587%

descriptions of places and people in narratives

No Comments

No Known Activations