INDEX

Explanations

words related to secrecy or unknown entities

oai_token-act-pair · gpt-3.5-turbo

references to things or concepts that are characterized as mysterious

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hers

-0.79

arf

-0.77

ffen

-0.76

acus

-0.75

mun

-0.73

onest

-0.69

zig

-0.69

Demand

-0.69

SN

-0.69

ighed

-0.69

POSITIVE LOGITS

 mysterious

0.90

 occurrences

0.89

 suspic

0.87

 disappearance

0.87

 unsolved

0.85

 unexplained

0.83

 coincidence

0.82

 circumstances

0.79

 landsl

0.78

worldly

0.76

Activations Density 0.018%

words related to secrecy or unknown entities

references to things or concepts that are characterized as mysterious

No Comments

No Known Activations