INDEX

Explanations

potential next steps or results

oai_token-act-pair · gpt-3.5-turbo

references to outcomes or results in various contexts

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

ker

-0.76

cer

-0.72

mens

-0.70

yi

-0.70

 brakes

-0.69

king

-0.69

ondo

-0.68

sis

-0.68

mans

-0.67

aps

-0.65

POSITIVE LOGITS

 outcome

1.11

 outcomes

0.99

bringer

0.83

 thereof

0.78

ebin

0.71

 result

0.68

fulness

0.67

 Cruel

0.65

aminer

0.65

 Orche

0.65

Activations Density 0.009%

potential next steps or results

references to outcomes or results in various contexts

No Comments

No Known Activations