INDEX

Explanations

numerical data, measurements, and technical terms related to records and readings

oai_token-act-pair · gpt-3.5-turbo

terms related to data collection and analysis, specifically in scientific or technical contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

sac

-0.81

capt

-0.76

rect

-0.73

arty

-0.71

olog

-0.70

struct

-0.69

vid

-0.69

ss

-0.68

bin

-0.68

 bystand

-0.68

POSITIVE LOGITS

 temperatures

0.96

 speeds

0.93

 Highest

0.90

 totals

0.88

ioxide

0.86

 exceeding

0.85

 estimate

0.82

 exceeded

0.81

 amount

0.80

 lows

0.80

Activations Density 0.171%

numerical data, measurements, and technical terms related to records and readings

terms related to data collection and analysis, specifically in scientific or technical contexts

No Comments

No Known Activations