INDEX

Explanations

the word "Ny" with various numerical suffixes attached to it

oai_token-act-pair · gpt-3.5-turbo

references to specific characters named "Nyssa" or related terms in a narrative context

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

eering

-0.89

ably

-0.85

OPLE

-0.80

Ø¯

-0.75

++++++++++++++++

-0.74

ÙĦ

-0.73

IBLE

-0.72

ãĥ´ãĤ¡

-0.69

Ø©

-0.69

urdue

-0.65

POSITIVE LOGITS

Ny

0.97

borg

0.88

heter

0.88

quist

0.87

ota

0.85

comed

0.84

seys

0.79

chel

0.76

yk

0.75

lus

0.74

Activations Density 0.016%

the word "Ny" with various numerical suffixes attached to it

references to specific characters named "Nyssa" or related terms in a narrative context

No Comments

No Known Activations