INDEX

Explanations

names

oai_token-act-pair · gpt-3.5-turbo

words related to specific names or identifiers, particularly related to people or potentially sensitive situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

fare

-0.73

hawks

-0.69

 cens

-0.64

pter

-0.63

mosp

-0.63

pron

-0.61

strings

-0.61

Elf

-0.61

 toile

-0.61

CES

-0.61

POSITIVE LOGITS

agate

3.12

angan

2.44

olen

1.60

asta

1.59

amon

1.35

olini

1.34

astern

1.29

oso

1.24

arella

1.17

atto

1.13

Activations Density 0.043%

names

words related to specific names or identifiers, particularly related to people or potentially sensitive situations

No Comments

No Known Activations