INDEX

Explanations

specific time-related events in a sequence of actions

oai_token-act-pair · gpt-3.5-turbo

words related to the beginning and ending of events or processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

tein

-0.69

anty

-0.68

wed

-0.67

 Inqu

-0.65

ardless

-0.64

 envy

-0.62

wing

-0.62

archives

-0.61

ãĥ¼ãĥ³

-0.58

Catholic

-0.58

POSITIVE LOGITS

 adjourn

0.73

 underway

0.72

 cris

0.65

 unfold

0.65

 imminent

0.64

undown

0.64

 dwind

0.64

orescence

0.61

iciary

0.60

 unfolding

0.60

Activations Density 0.253%

specific time-related events in a sequence of actions

words related to the beginning and ending of events or processes

No Comments

No Known Activations