INDEX

Explanations

the word "regain" or its variations

oai_token-act-pair · gpt-3.5-turbo

terms related to regaining control or status

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

QB

-0.64

cas

-0.64

tar

-0.64

ahi

-0.63

mis

-0.63

gom

-0.63

raid

-0.62

abad

-0.61

picture

-0.59

California

-0.58

POSITIVE LOGITS

 regain

1.24

 regained

1.09

chwitz

0.86

imar

0.86

overy

0.82

 recovery

0.81

 braking

0.81

 recapt

0.81

 composure

0.81

 control

0.78

Activations Density 0.012%

the word "regain" or its variations

terms related to regaining control or status

No Comments

No Known Activations