INDEX

Explanations

references to people's names, especially the repeated mention of "Ruth" and "Babe Ruth"

oai_token-act-pair · gpt-3.5-turbo

references to the name "Ruth."

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

gotten

-0.73

ctica

-0.67

agons

-0.62

olesc

-0.62

artney

-0.62

iator

-0.62

tein

-0.61

 Helsinki

-0.61

opal

-0.61

akening

-0.60

POSITIVE LOGITS

anne

0.86

 Ruth

0.85

lessly

0.84

uth

0.79

less

0.77

lessness

0.74

enthal

0.71

utherford

0.71

mite

0.70

anna

0.68

Activations Density 0.012%

references to people's names, especially the repeated mention of "Ruth" and "Babe Ruth"

references to the name "Ruth."

No Comments

No Known Activations