INDEX

Explanations

information about actions being taken to address specific issues or goals

oai_token-act-pair · gpt-3.5-turbo

phrases indicating actions or measures being taken

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.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

 ç¥ŀ

-0.70

olls

-0.64

itability

-0.64

 slic

-0.61

ibble

-0.61

irl

-0.61

 Prediction

-0.61

parts

-0.61

Osc

-0.61

anny

-0.61

POSITIVE LOGITS

hooting

0.93

 measures

0.90

 precautions

0.85

 implemented

0.84

 remed

0.83

 steps

0.82

 undertaken

0.81

 backward

0.79

 safeguards

0.78

 actions

0.78

Activations Density 0.102%

information about actions being taken to address specific issues or goals

phrases indicating actions or measures being taken

No Comments

No Known Activations