INDEX

Explanations

terms related to improvisation and development

oai_token-act-pair · gpt-3.5-turbo

terms related to improvisation and performance

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

holders

-0.87

hips

-0.83

houses

-0.83

iard

-0.73

track

-0.68

metics

-0.68

roid

-0.68

 Blackwell

-0.68

rooms

-0.68

maps

-0.67

POSITIVE LOGITS

 improv

1.09

 improvis

0.99

bably

0.90

atively

0.88

 Impro

0.81

 comedy

0.79

vernment

0.78

itone

0.75

idently

0.74

 rehears

0.73

Activations Density 0.020%

terms related to improvisation and development

terms related to improvisation and performance

No Comments

No Known Activations