INDEX

Explanations

questions and statements addressing the reader or listener directly

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

yrights

-0.83

ces

-0.78

inery

-0.73

edIn

-0.72

opens

-0.72

ooks

-0.71

ges

-0.69

ruciating

-0.68

tails

-0.68

uces

-0.68

POSITIVE LOGITS

?'

1.06

?'"

1.04

 ever

1.01

?"

0.98

0.95

?)

0.94

?:

0.92

...?

0.90

?!

0.86

????

0.86

Activations Density 0.513%

questions and statements addressing the reader or listener directly

No Comments

No Known Activations