INDEX

Explanations

references to government figures and positions in political contexts

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

 goddamn

-0.84

 darn

-0.77

....

-0.75

 damn

-0.74

etc

-0.74

().

-0.70

..."

-0.69

 somet

-0.69

$.

-0.69

 ..."

-0.67

POSITIVE LOGITS

 spokesman

0.80

 spokeswoman

0.77

 lawmakers

0.73

 meanwhile

0.71

 campaigners

0.71

 Reuters

0.70

 Hollande

0.70

resa

0.69

 lawmaker

0.69

 diplomats

0.67

Activations Density 0.841%

references to government figures and positions in political contexts

No Comments

No Known Activations