INDEX

Explanations

occurrences of specific religious titles and terms

oai_token-act-pair · gpt-3.5-turbo

mentions of religious figures and texts, particularly related to Judaism

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

swick

-0.87

 Hurricanes

-0.85

mson

-0.77

oples

-0.73

iago

-0.72

eworld

-0.71

auga

-0.71

 Gators

-0.71

umbnail

-0.68

osure

-0.67

POSITIVE LOGITS

 rabb

1.14

 rabbi

1.05

 Rabbi

0.98

×ľ

0.96

 Torah

0.95

anyahu

0.93

×ª

0.92

×

0.90

×©

0.89

×ķ

0.89

Activations Density 0.041%

occurrences of specific religious titles and terms

mentions of religious figures and texts, particularly related to Judaism

No Comments

No Known Activations