INDEX

Explanations

objects related to events or actions that have occurred.

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "that" in various grammatical contexts

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

 VIDEOS

-0.71

ivating

-0.68

emonic

-0.66

 Unlimited

-0.63

 Checking

-0.60

 Expect

-0.60

 adren

-0.60

 guessing

-0.59

 Riding

-0.59

EngineDebug

-0.59

POSITIVE LOGITS

 belonged

1.19

 resembled

0.96

 resided

0.93

 housed

0.92

 protr

0.91

soever

0.90

 belongs

0.88

 resulted

0.83

 escaped

0.83

 weighs

0.82

Activations Density 0.234%

objects related to events or actions that have occurred.

occurrences of the word "that" in various grammatical contexts

No Comments

No Known Activations