INDEX

Explanations

terms related to a logical or situational contradiction

oai_token-act-pair · gpt-3.5-turbo

references to paradoxes and contradictions

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

Interstitial

-0.86

rations

-0.68

baugh

-0.67

ahon

-0.66

psc

-0.65

igree

-0.64

collar

-0.64

 ballpark

-0.64

mails

-0.63

estones

-0.63

POSITIVE LOGITS

 paradox

1.12

 Paradox

1.06

ical

1.01

ioned

0.95

undrum

0.93

naire

0.83

 Despair

0.80

otomy

0.77

ically

0.76

aires

0.75

Activations Density 0.025%

terms related to a logical or situational contradiction

references to paradoxes and contradictions

No Comments

No Known Activations