INDEX

Explanations

information about events or processes happening before a certain starting point

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate summaries or evaluations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ãĥŁ

-0.89

ulla

-0.71

atre

-0.69

mite

-0.68

ahu

-0.68

è£ħ

-0.68

ãĤ¶

-0.64

æĺ

-0.64

ãĤ§

-0.63

Charges

-0.62

POSITIVE LOGITS

how

1.31

 including

1.05

 whether

1.04

 what

0.98

 namely

0.96

why

0.96

how

0.93

 comparing

0.91

 ranging

0.91

 noting

0.86

Activations Density 0.252%

information about events or processes happening before a certain starting point

phrases that indicate summaries or evaluations

No Comments

No Known Activations