INDEX

Explanations

media mentions of television networks

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

 signatures

-0.74

 dormant

-0.69

pta

-0.68

 folds

-0.67

 extinct

-0.67

artifacts

-0.66

ger

-0.64

 tuber

-0.63

 ferment

-0.63

 swelling

-0.63

POSITIVE LOGITS

radio

1.16

levision

1.13

Radio

1.12

NBC

1.11

 radio

1.04

NBC

1.04

CBS

1.00

ABC

0.99

TV

0.99

FOX

0.99

Activations Density 0.084%

media mentions of television networks

No Comments

No Known Activations