INDEX

Explanations

terms related to canonical texts or entities

oai_token-act-pair · gpt-3.5-turbo

terms related to religious texts and their authoritative frameworks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

berman

-0.76

bler

-0.75

ez

-0.72

hard

-0.72

ld

-0.71

hani

-0.69

æĸ¹

-0.67

ELL

-0.67

sol

-0.67

activity

-0.67

POSITIVE LOGITS

icals

1.02

istical

0.93

ical

0.90

hedral

0.88

icity

0.88

ICAL

0.87

ization

0.84

ieve

0.83

ize

0.82

ique

0.80

Activations Density 0.089%

terms related to canonical texts or entities

terms related to religious texts and their authoritative frameworks

No Comments

No Known Activations