INDEX

Explanations

names or words related to specific individuals

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or entities in a political context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

sburgh

-0.83

sers

-0.75

sburg

-0.74

IVERS

-0.73

afia

-0.71

namese

-0.69

 behavi

-0.68

 spirited

-0.67

ersion

-0.67

ainment

-0.66

POSITIVE LOGITS

arella

0.89

eye

0.77

hao

0.75

aman

0.75

phrine

0.75

agame

0.74

orah

0.74

ipped

0.73

eta

0.72

arks

0.72

Activations Density 0.024%

names or words related to specific individuals

references to specific individuals or entities in a political context

No Comments

No Known Activations