INDEX

Explanations

words related to administrative actions and processes

oai_token-act-pair · gpt-3.5-turbo

references to administrative matters or processes

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

IFrame

Link

Negative Logits

eem

-0.91

ï¸

-0.87

hov

-0.84

âķĲâķĲ

-0.81

ku

-0.78

ï¸ı

-0.76

ooth

-0.75

KI

-0.75

pots

-0.72

kok

-0.71

POSITIVE LOGITS

 procedures

0.91

rative

0.90

 administrative

0.87

 substitution

0.83

 tasks

0.82

 disciplinary

0.82

 instit

0.82

 procedure

0.81

izational

0.81

 administrator

0.80

Activations Density 0.018%

words related to administrative actions and processes

references to administrative matters or processes

No Known Activations