INDEX

Explanations

phrases related to anonymity or conditions set for certain actions

oai_token-act-pair · gpt-3.5-turbo

references to conditions or stipulations in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

encer

-0.82

riet

-0.69

umbn

-0.67

 Thro

-0.66

 erupted

-0.64

encers

-0.63

 Highlander

-0.61

rious

-0.59

urity

-0.58

 Springfield

-0.58

POSITIVE LOGITS

 confir

0.87

 elabor

0.79

 indemn

0.75

 condition

0.74

Condition

0.72

 caveat

0.72

 lett

0.70

 preserving

0.69

atives

0.69

 agreeing

0.69

Activations Density 0.028%

phrases related to anonymity or conditions set for certain actions

references to conditions or stipulations in various contexts

No Comments

No Known Activations