INDEX

Explanations

mentions of family members, particularly sons

oai_token-act-pair · gpt-3.5-turbo

references to "son" 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 @ 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

veyard

-0.82

iculty

-0.78

kefeller

-0.73

 proport

-0.63

orney

-0.63

pmwiki

-0.62

 manifold

-0.61

topic

-0.61

PORT

-0.61

Population

-0.61

POSITIVE LOGITS

hood

1.00

 Gohan

0.98

hesis

0.89

nets

0.86

heses

0.83

pins

0.82

ogram

0.82

son

0.80

friend

0.78

 Barron

0.77

Activations Density 0.019%

mentions of family members, particularly sons

references to "son" in various contexts

No Comments

No Known Activations