INDEX

Explanations

names of locations or people, specifically those with unique spellings or structures, such as "Higbee" or "Kwabena Boateng"

oai_token-act-pair · gpt-3.5-turbo

names of specific individuals or key figures

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

ãĥ¼ãĥĨ

-0.85

 Bundy

-0.72

Â¯Â¯

-0.69

ãĥĵ

-0.69

 Savior

-0.68

 Dynamics

-0.67

âĸ¬âĸ¬

-0.67

ãĤ´ãĥ³

-0.65

ãĥĩãĤ£

-0.65

 Primordial

-0.64

POSITIVE LOGITS

imura

0.97

umen

0.96

imon

0.89

umin

0.88

una

0.88

oku

0.86

raf

0.85

alach

0.84

unda

0.84

inski

0.83

Activations Density 0.032%

names of locations or people, specifically those with unique spellings or structures, such as "Higbee" or "Kwabena Boateng"

names of specific individuals or key figures

No Comments

No Known Activations