INDEX

Explanations

proper names

oai_token-act-pair · gpt-3.5-turbo

proper nouns, especially names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

-0.82

scenes

-0.73

birds

-0.69

 Tactics

-0.68

cells

-0.68

mallow

-0.67

bands

-0.66

 sled

-0.65

clad

-0.62

 Conduct

-0.62

POSITIVE LOGITS

igible

0.85

Edu

0.83

undo

0.81

estine

0.80

aughed

0.80

ardo

0.78

ston

0.77

rative

0.76

ucer

0.74

itialized

0.73

Activations Density 0.029%

proper names

proper nouns, especially names

No Comments

No Known Activations