INDEX

Explanations

references to FBI and related activities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

piring

-0.72

faces

-0.71

ijk

-0.70

pain

-0.68

flush

-0.68

mere

-0.68

ktop

-0.67

criptions

-0.67

eem

-0.67

lighting

-0.66

POSITIVE LOGITS

 informant

0.98

 investigation

0.94

 counterterrorism

0.93

 Agents

0.91

 Director

0.91

 Inspector

0.86

 Investigations

0.86

 Headquarters

0.85

 Investigation

0.85

 Bureau

0.84

Activations Density 0.025%

references to FBI and related activities

No Comments

No Known Activations