INDEX

Explanations

words related to contrasting or opposing ideas

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ires

-0.77

estamp

-0.63

DERR

-0.60

bourg

-0.59

ascus

-0.59

ptoms

-0.58

ested

-0.57

tackle

-0.57

ernels

-0.56

erest

-0.56

POSITIVE LOGITS

ifice

1.18

Else

1.16

acle

1.13

acles

1.13

chard

1.04

acular

1.02

chid

1.01

nery

0.98

 alternatively

0.98

lando

0.97

Activations Density 0.115%

words related to contrasting or opposing ideas

No Comments

No Known Activations

words related to contrasting or opposing ideas

No Comments

No Known Activations