INDEX

Explanations

words related to criticism or issues related to being flawed or faulty

oai_token-act-pair · gpt-3.5-turbo

terms related to flaws and defects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ciation

-0.87

iser

-0.73

ista

-0.72

VO

-0.71

Shop

-0.70

isation

-0.70

 Colleges

-0.69

ervation

-0.67

 COURT

-0.66

ILA

-0.65

POSITIVE LOGITS

 flawed

1.19

 compromises

0.91

 flaws

0.84

 flaw

0.81

matically

0.76

 caric

0.76

 guiActiveUn

0.73

 unreliable

0.72

glers

0.71

 failures

0.70

Activations Density 0.012%

words related to criticism or issues related to being flawed or faulty

terms related to flaws and defects

No Comments

No Known Activations