INDEX

Explanations

films and TV shows that are based on existing works or real-life events

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

bats

-0.86

vous

-0.81

limit

-0.75

fax

-0.74

zon

-0.70

ura

-0.70

asive

-0.70

aii

-0.69

©¶æ¥µ

-0.69

xus

-0.69

POSITIVE LOGITS

 principles

0.84

 scratch

0.84

 combining

0.82

 principle

0.77

 collaboration

0.75

 necessity

0.74

 scraps

0.73

 belief

0.73

 conjecture

0.73

 proprietary

0.72

Activations Density 1.460%

films and TV shows that are based on existing works or real-life events

No Comments

No Known Activations