INDEX

Explanations

terms related to incorrect beliefs or understanding

oai_token-act-pair · gpt-3.5-turbo

terms related to misconceptions and misinformation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

igree

-0.74

atom

-0.72

negie

-0.71

ramid

-0.69

edom

-0.67

imen

-0.66

itar

-0.66

incinn

-0.66

ropri

-0.65

gans

-0.65

POSITIVE LOGITS

 misunderstanding

1.01

 misconceptions

1.00

 misconception

0.94

 misunderstand

0.87

 inaccur

0.85

 misinformation

0.84

 misinterpret

0.84

 incorrectly

0.81

 misunderstood

0.77

 inaccurate

0.76

Activations Density 0.044%

terms related to incorrect beliefs or understanding

terms related to misconceptions and misinformation

No Comments

No Known Activations