INDEX

Explanations

references to personal experiences and conversations, especially those that convey community interaction or discussions over the years

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

oret

-0.74

 Hezbollah

-0.67

elfare

-0.65

-0.63

Assad

-0.62

rous

-0.61

 Clause

-0.61

ificantly

-0.61

 Kissinger

-0.60

 Hamas

-0.60

POSITIVE LOGITS

 blogging

1.06

 myself

1.02

 hobby

0.95

 haha

0.88

 browsing

0.87

 hobbies

0.85

 researching

0.84

 undergrad

0.79

 geek

0.78

 homebrew

0.78

Activations Density 0.553%

references to personal experiences and conversations, especially those that convey community interaction or discussions over the years

No Comments

No Known Activations