INDEX

Explanations

phrases related to time or sequence, specifically focusing on occurrences happening earlier

oai_token-act-pair · gpt-3.5-turbo

references to time, specifically instances of the word "earlier" indicating past events

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

Bal

-0.60

Pros

-0.60

Optional

-0.58

Rust

-0.57

esome

-0.57

odor

-0.54

eus

-0.54

pmwiki

-0.53

especially

-0.52

 Skydragon

-0.52

POSITIVE LOGITS

 this

1.11

this

0.93

 today

0.93

 than

0.91

 tonight

0.87

than

0.87

 THIS

0.82

 Thursday

0.74

today

0.73

stage

0.73

Activations Density 0.042%

phrases related to time or sequence, specifically focusing on occurrences happening earlier

references to time, specifically instances of the word "earlier" indicating past events

No Comments

No Known Activations