INDEX

Explanations

actions related to accountability

oai_token-act-pair · gpt-3.5-turbo

terms related to legal and procedural actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ãĥİ

-0.79

æ©Ł

-0.79

orld

-0.74

FORE

-0.68

iuses

-0.68

ãĥ¼ãĥĨãĤ£

-0.67

 NEVER

-0.67

 teasp

-0.66

emis

-0.64

Discussion

-0.63

POSITIVE LOGITS

 them

1.07

enance

0.96

him

0.91

 theirs

0.82

 such

0.80

 their

0.80

any

0.78

 oneself

0.77

 runaway

0.76

 anything

0.76

Activations Density 0.458%

actions related to accountability

terms related to legal and procedural actions

No Comments

No Known Activations