INDEX

Explanations

arguments and claims related to misconceptions, falsehoods, and controversies

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

dos

-0.78

artney

-0.77

ktop

-0.75

Interstitial

-0.74

ophon

-0.68

 actionGroup

-0.67

incinn

-0.66

zie

-0.65

mes

-0.64

eteria

-0.64

POSITIVE LOGITS

 assertions

1.15

 assertion

1.13

 debunked

1.12

 belief

1.06

icist

1.06

 validity

1.04

 assumptions

1.03

 untrue

1.02

 falsehood

0.99

 debunk

0.97

Activations Density 5.079%

arguments and claims related to misconceptions, falsehoods, and controversies

No Comments

No Known Activations