INDEX

Explanations

references to actions or verbs that indicate direction or movement

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

clerosis

-0.73

Roy

-0.71

deg

-0.70

ounded

-0.69

DEV

-0.68

tu

-0.66

Bay

-0.66

ģ«

-0.66

independence

-0.66

azel

-0.64

POSITIVE LOGITS

guy

0.84

 solution

0.82

 superpower

0.81

 destination

0.80

 recommendation

0.79

 option

0.77

 babys

0.77

 remedy

0.76

 conduit

0.76

 culprit

0.73

Activations Density 0.011%

references to actions or verbs that indicate direction or movement

No Comments

No Known Activations