INDEX

Explanations

information related to events, locations, and activities

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

soType

-0.27

ãĤ§

-0.26

Cause

-0.24

iability

-0.24

matter

-0.24

hesion

-0.24

independence

-0.24

agn

-0.23

arn

-0.23

oxin

-0.22

POSITIVE LOGITS

TWO

0.25

CoC

0.24

 Dunk

0.24

WW

0.23

 yesterday

0.23

 Quote

0.23

 actual

0.22

 tonight

0.22

 Recall

0.22

 nomine

0.22

Activations Density 15.290%

information related to events, locations, and activities

No Comments

No Known Activations