INDEX

Explanations

proper nouns related to individuals or places

oai_token-act-pair · gpt-3.5-turbo

references to characters or names associated with specific films or stories

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

 Contra

-0.79

ãĥ¼ãĥĨ

-0.77

ãĥĩãĤ£

-0.74

 Fargo

-0.66

 CLSID

-0.64

odcast

-0.62

 stump

-0.61

 Tome

-0.60

«ĺ

-0.60

 Dragonbound

-0.60

POSITIVE LOGITS

affe

1.05

aval

0.86

wx

0.81

oux

0.79

abus

0.78

aldi

0.77

otta

0.76

onna

0.75

bons

0.74

acha

0.74

Activations Density 0.029%

proper nouns related to individuals or places

references to characters or names associated with specific films or stories

No Comments

No Known Activations