INDEX

Explanations

names related to people

oai_token-act-pair · gpt-3.5-turbo

the word "lo" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ICLE

-0.74

ãĥĩ

-0.71

chell

-0.67

glass

-0.66

 Equality

-0.64

 Dispatch

-0.64

 Democr

-0.64

imental

-0.63

 provoking

-0.62

icles

-0.62

POSITIVE LOGITS

fty

1.17

zzle

1.15

zzi

1.09

pper

1.06

ppy

1.03

opy

0.96

veland

0.94

pped

0.93

oyd

0.92

zz

0.91

Activations Density 0.007%

names related to people

the word "lo" in various contexts

No Comments

No Known Activations