INDEX

Explanations

terms related to campaigns, specifically those that are political or environmental in nature

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

ortment

-0.69

hap

-0.66

ometimes

-0.65

ahime

-0.64

 Bowen

-0.62

 Listen

-0.60

enhagen

-0.60

eele

-0.60

rongh

-0.59

Fig

-0.59

POSITIVE LOGITS

 whatsoever

1.61

nor

1.33

 anymore

1.07

 slightest

0.96

 except

0.87

 requirement

0.85

nor

0.85

 markings

0.82

 avail

0.81

 anywhere

0.79

Activations Density 0.158%

terms related to campaigns, specifically those that are political or environmental in nature

No Comments

No Known Activations