INDEX

Explanations

dates written in a specific format (MM/DD/YYYY)

oai_token-act-pair · gpt-3.5-turbo

occurrences of numerical values and their contextual references

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

foc

-0.93

 favour

-0.82

 appe

-0.79

 analys

-0.79

 aggreg

-0.78

 incub

-0.77

 organis

-0.76

 seiz

-0.75

 stabil

-0.75

 regener

-0.75

POSITIVE LOGITS

WASHINGTON

2.17

SAN

1.97

LOS

1.88

NEW

1.83

GREEN

1.71

DEN

1.68

COL

1.66

CLE

1.66

HOU

1.65

CHAR

1.64

Activations Density 0.140%

dates written in a specific format (MM/DD/YYYY)

occurrences of numerical values and their contextual references

No Comments

No Known Activations