INDEX

Explanations

statements related to truth and falsehood

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

 Nanto

-0.75

aths

-0.72

 renovations

-0.72

 Chains

-0.71

 UCHIJ

-0.71

orers

-0.70

 contrace

-0.70

iencies

-0.69

ching

-0.69

 downtime

-0.68

POSITIVE LOGITS

 unfounded

1.32

 baseless

1.29

 untrue

1.13

 refuted

1.10

 valid

1.06

 debunked

1.04

 false

1.03

 contradicted

1.00

 nonsense

1.00

 incorrect

0.99

Activations Density 0.345%

statements related to truth and falsehood

No Comments

No Known Activations