INDEX

Explanations

instances of information being deemed incorrect or inaccurate

oai_token-act-pair · gpt-3.5-turbo

terms that denote falsehood or inaccuracies

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.2.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

agra

-0.99

abiding

-0.79

ramid

-0.78

trak

-0.77

neys

-0.76

mun

-0.75

hibit

-0.73

aea

-0.73

gans

-0.72

hov

-0.72

POSITIVE LOGITS

 inaccurate

1.11

 inaccur

1.10

 incorrect

1.04

 erroneous

0.99

 misinformation

0.98

 guiActiveUn

0.96

 misconceptions

0.96

 incorrectly

0.93

 mistaken

0.90

 misconception

0.90

Activations Density 0.028%

instances of information being deemed incorrect or inaccurate

terms that denote falsehood or inaccuracies

No Comments

No Known Activations