INDEX

Explanations

activating phrases related to moral or ethical judgments regarding actions or beliefs

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

bour

-0.61

rament

-0.56

afa

-0.56

 Dynam

-0.54

 Horizons

-0.53

ordon

-0.53

Chapter

-0.52

wer

-0.51

Ack

-0.49

asher

-0.48

POSITIVE LOGITS

to

1.17

to

1.12

To

0.90

 thereto

0.89

eto

0.87

TO

0.86

TO

0.84

ta

0.80

onte

0.79

 Nanto

0.79

Activations Density 0.156%

activating phrases related to moral or ethical judgments regarding actions or beliefs

No Comments

No Known Activations