INDEX

Explanations

recently occurring events or trends

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "lately" and its context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.97

ammy

-0.81

aughtered

-0.73

eer

-0.71

IAL

-0.69

tained

-0.67

pperc

-0.66

atively

-0.65

race

-0.65

ultural

-0.64

POSITIVE LOGITS

 lately

1.26

 inactive

0.85

 overlooked

0.81

inventoryQuantity

0.75

 slump

0.75

 antioxid

0.73

 unsus

0.73

 overshadowed

0.72

bc

0.71

 retreating

0.71

Activations Density 0.005%

recently occurring events or trends

occurrences of the word "lately" and its context

No Comments

No Known Activations