INDEX

Explanations

statements or reports released by individuals or organizations

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

uncture

-0.73

ampires

-0.71

 cared

-0.67

 belongs

-0.66

 afforded

-0.65

 Played

-0.64

eno

-0.63

 belonged

-0.63

rex

-0.63

builders

-0.63

POSITIVE LOGITS

 statement

1.16

 scathing

1.13

 memorandum

1.04

 preliminary

1.03

 rebutt

1.01

 apology

0.99

 report

0.98

 condolences

0.98

 condol

0.98

 Statement

0.97

Activations Density 1.647%

statements or reports released by individuals or organizations

No Comments

No Known Activations