INDEX

Explanations

terms related to risk and strength metrics

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

SPONSORED

-0.98

alter

-0.83

IGHTS

-0.81

=-=-=-=-

-0.75

umi

-0.72

adier

-0.71

ynthesis

-0.71

agi

-0.70

$$$$

-0.67

ãĤ¨ãĥ«

-0.67

POSITIVE LOGITS

(>

1.08

 earners

0.79

(<

0.76

 areas

0.75

 interval

0.71

 reactive

0.70

 urban

0.69

 residential

0.68

 airborne

0.68

 environments

0.67

Activations Density 0.073%

terms related to risk and strength metrics

No Comments

No Known Activations