INDEX

Explanations

names related to influential individuals

oai_token-act-pair · gpt-3.5-turbo

proper names, particularly those of individuals associated with specific actions or 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

éĹĺ

-0.98

izations

-0.82

heast

-0.82

ufact

-0.78

xual

-0.77

psons

-0.77

arily

-0.74

 coup

-0.74

redits

-0.74

mingham

-0.72

POSITIVE LOGITS

Hub

0.87

iasis

0.77

 Malfoy

0.71

asar

0.69

 Feinstein

0.68

adium

0.68

aho

0.67

 glare

0.67

Wan

0.67

 Sheldon

0.67

Activations Density 0.035%

names related to influential individuals

proper names, particularly those of individuals associated with specific actions or contexts

No Comments

No Known Activations