INDEX

Explanations

references to various individuals' names

oai_token-act-pair · gpt-3.5-turbo

instances of the word "om."

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

LIST

-0.69

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

-0.66

 convict

-0.62

 nexus

-0.61

 separatist

-0.59

 fact

-0.59

REDACTED

-0.59

 Kindle

-0.58

 Brees

-0.57

Wad

-0.57

POSITIVE LOGITS

puter

1.14

orrow

1.14

obile

1.13

otive

1.07

atoes

1.05

useum

1.02

obiles

0.98

otor

0.96

pson

0.96

ikawa

0.96

Activations Density 0.013%

references to various individuals' names

instances of the word "om."

No Comments

No Known Activations