INDEX

Explanations

expressions of hope or uncertainty regarding future developments or plans

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

 dehuman

-0.67

 pearl

-0.61

odied

-0.61

 misogyn

-0.60

 discriminated

-0.60

 indistinguishable

-0.58

 Godd

-0.57

seless

-0.56

 falsely

-0.56

 vulgar

-0.55

POSITIVE LOGITS

 timetable

0.95

0.89

TBD

0.86

 meantime

0.86

catentry

0.83

 hopefully

0.82

Hopefully

0.81

 deadlines

0.80

 deadline

0.80

TBA

0.77

Activations Density 0.665%

expressions of hope or uncertainty regarding future developments or plans

No Comments

No Known Activations